What It Was
During the COVID-19 pandemic, Michigan residents needed access to a range of emergency assistance programs — unemployment benefits, rental assistance, food aid, and more. The existing state infrastructure wasn't built for this volume or speed.
I was part of the team that built the public-facing applications that allowed Michigan residents to apply for, check the status of, and manage their assistance cases.
The Problem It Solved
The existing systems were slow, phone-based, and overwhelmed. People were waiting days or weeks to find out if they'd get help they needed immediately. The state needed software fast, and it needed to actually work for people under extreme stress.
"Actually work" meant: load on a cracked phone screen, operate on spotty data, handle multiple languages, and output clean data that fed into government benefit systems — while the government systems themselves were being reconfigured on the fly.
What I Built
- Public-facing application flows for unemployment and rental assistance programs
- Case status tracking for applicants
- Integration layer connecting application submissions to state eligibility systems
- Admin tools for state employees to process and manage cases at volume
The Outcome
Hundreds of thousands of Michigan residents used these applications to access emergency relief during the hardest stretch of their lives. The apps handled the load. The integrations worked. People got help.
That's the whole metric. That's the one that matters.
What This Project Made Clear
This was the project that taught me what high-stakes software really means. Not stakes in terms of revenue. Stakes in terms of whether someone can pay rent this month.
Everything since has been filtered through that. If it doesn't actually work for real people in real conditions, it didn't ship.